webapp.py 11 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346
  1. #!/usr/bin/env python
  2. '''
  3. searx is free software: you can redistribute it and/or modify
  4. it under the terms of the GNU Affero General Public License as published by
  5. the Free Software Foundation, either version 3 of the License, or
  6. (at your option) any later version.
  7. searx is distributed in the hope that it will be useful,
  8. but WITHOUT ANY WARRANTY; without even the implied warranty of
  9. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  10. GNU Affero General Public License for more details.
  11. You should have received a copy of the GNU Affero General Public License
  12. along with searx. If not, see < http://www.gnu.org/licenses/ >.
  13. (C) 2013- by Adam Tauber, <asciimoo@gmail.com>
  14. '''
  15. import json
  16. import cStringIO
  17. import os
  18. from flask import Flask, request, render_template
  19. from flask import url_for, Response, make_response, redirect
  20. from flask import send_from_directory
  21. from searx import settings, searx_dir
  22. from searx.engines import search, categories, engines, get_engines_stats
  23. from searx.utils import UnicodeWriter
  24. from searx.utils import highlight_content, html_to_text
  25. from searx.languages import language_codes
  26. from flask.ext.babel import Babel
  27. app = Flask(
  28. __name__,
  29. static_folder=os.path.join(searx_dir, 'static'),
  30. template_folder=os.path.join(searx_dir, 'templates')
  31. )
  32. app.secret_key = settings['server']['secret_key']
  33. babel = Babel(app)
  34. #TODO configurable via settings.yml
  35. favicons = ['wikipedia', 'youtube', 'vimeo', 'soundcloud',
  36. 'twitter', 'stackoverflow', 'github']
  37. @babel.localeselector
  38. def get_locale():
  39. locale = request.accept_languages.best_match(settings['locales'].keys())
  40. if request.cookies.get('locale', '') in settings['locales']:
  41. locale = request.cookies.get('locale', '')
  42. if 'locale' in request.args\
  43. and request.args['locale'] in settings['locales']:
  44. locale = request.args['locale']
  45. if 'locale' in request.form\
  46. and request.form['locale'] in settings['locales']:
  47. locale = request.form['locale']
  48. return locale
  49. def get_base_url():
  50. if settings['server']['base_url']:
  51. hostname = settings['server']['base_url']
  52. else:
  53. scheme = 'http'
  54. if request.is_secure:
  55. scheme = 'https'
  56. hostname = url_for('index', _external=True, _scheme=scheme)
  57. return hostname
  58. def render(template_name, **kwargs):
  59. global categories
  60. kwargs['categories'] = ['general']
  61. kwargs['categories'].extend(x for x in
  62. sorted(categories.keys()) if x != 'general')
  63. if not 'selected_categories' in kwargs:
  64. kwargs['selected_categories'] = []
  65. cookie_categories = request.cookies.get('categories', '').split(',')
  66. for ccateg in cookie_categories:
  67. if ccateg in categories:
  68. kwargs['selected_categories'].append(ccateg)
  69. if not len(kwargs['selected_categories']):
  70. kwargs['selected_categories'] = ['general']
  71. return render_template(template_name, **kwargs)
  72. def parse_query(query):
  73. query_engines = []
  74. query_parts = query.split()
  75. if query_parts[0].startswith('-'):
  76. engine_name = query_parts[0][1:].replace('_', ' ')
  77. if engine_name in engines:
  78. query_engines.append({'category': 'none',
  79. 'name': query_parts[0][1:]})
  80. elif query_parts[0].startswith('!'):
  81. category_name = query_parts[0][1:].replace('_', ' ')
  82. if category_name in categories:
  83. query_engines.extend({'category': category_name,
  84. 'name': engine.name}
  85. for engine in categories[category_name])
  86. if len(query_engines):
  87. query = query.replace(query_parts[0], '', 1).strip()
  88. return query, query_engines
  89. @app.route('/', methods=['GET', 'POST'])
  90. def index():
  91. paging = False
  92. lang = 'all'
  93. if request.cookies.get('language')\
  94. and request.cookies['language'] in (x[0] for x in language_codes):
  95. lang = request.cookies['language']
  96. if request.method == 'POST':
  97. request_data = request.form
  98. else:
  99. request_data = request.args
  100. if not request_data.get('q'):
  101. return render('index.html')
  102. pageno_param = request_data.get('pageno', '1')
  103. if not pageno_param.isdigit() or int(pageno_param) < 1:
  104. return render('index.html')
  105. pageno = int(pageno_param)
  106. selected_categories = []
  107. query, selected_engines = parse_query(request_data['q'].encode('utf-8'))
  108. if len(selected_engines):
  109. selected_categories = list(set(engine['category']
  110. for engine in selected_engines))
  111. else:
  112. for pd_name, pd in request_data.items():
  113. if pd_name.startswith('category_'):
  114. category = pd_name[9:]
  115. if not category in categories:
  116. continue
  117. selected_categories.append(category)
  118. if not len(selected_categories):
  119. cookie_categories = request.cookies.get('categories', '')
  120. cookie_categories = cookie_categories.split(',')
  121. for ccateg in cookie_categories:
  122. if ccateg in categories:
  123. selected_categories.append(ccateg)
  124. if not len(selected_categories):
  125. selected_categories = ['general']
  126. for categ in selected_categories:
  127. selected_engines.extend({'category': categ,
  128. 'name': x.name}
  129. for x in categories[categ])
  130. results, suggestions = search(query,
  131. request,
  132. selected_engines,
  133. pageno,
  134. lang)
  135. for result in results:
  136. if not paging and engines[result['engine']].paging:
  137. paging = True
  138. if request_data.get('format', 'html') == 'html':
  139. if 'content' in result:
  140. result['content'] = highlight_content(result['content'], query)
  141. result['title'] = highlight_content(result['title'], query)
  142. else:
  143. if 'content' in result:
  144. result['content'] = html_to_text(result['content']).strip()
  145. result['title'] = html_to_text(result['title']).strip()
  146. if len(result['url']) > 74:
  147. url_parts = result['url'][:35], result['url'][-35:]
  148. result['pretty_url'] = '{0}[...]{1}'.format(*url_parts)
  149. else:
  150. result['pretty_url'] = result['url']
  151. for engine in result['engines']:
  152. if engine in favicons:
  153. result['favicon'] = engine
  154. if request_data.get('format') == 'json':
  155. return Response(json.dumps({'query': query, 'results': results}),
  156. mimetype='application/json')
  157. elif request_data.get('format') == 'csv':
  158. csv = UnicodeWriter(cStringIO.StringIO())
  159. keys = ('title', 'url', 'content', 'host', 'engine', 'score')
  160. if len(results):
  161. csv.writerow(keys)
  162. for row in results:
  163. row['host'] = row['parsed_url'].netloc
  164. csv.writerow([row.get(key, '') for key in keys])
  165. csv.stream.seek(0)
  166. response = Response(csv.stream.read(), mimetype='application/csv')
  167. content_disp = 'attachment;Filename=searx_-_{0}.csv'.format(query)
  168. response.headers.add('Content-Disposition', content_disp)
  169. return response
  170. elif request_data.get('format') == 'rss':
  171. response_rss = render(
  172. 'opensearch_response_rss.xml',
  173. results=results,
  174. q=request_data['q'],
  175. number_of_results=len(results),
  176. base_url=get_base_url()
  177. )
  178. return Response(response_rss, mimetype='text/xml')
  179. return render(
  180. 'results.html',
  181. results=results,
  182. q=request_data['q'],
  183. selected_categories=selected_categories,
  184. paging=paging,
  185. pageno=pageno,
  186. suggestions=suggestions
  187. )
  188. @app.route('/about', methods=['GET'])
  189. def about():
  190. return render('about.html')
  191. @app.route('/engines', methods=['GET'])
  192. def list_engines():
  193. global categories
  194. return render('engines.html', categs=categories.items())
  195. @app.route('/preferences', methods=['GET', 'POST'])
  196. def preferences():
  197. lang = None
  198. if request.cookies.get('language')\
  199. and request.cookies['language'] in (x[0] for x in language_codes):
  200. lang = request.cookies['language']
  201. if request.method == 'POST':
  202. selected_categories = []
  203. locale = None
  204. for pd_name, pd in request.form.items():
  205. if pd_name.startswith('category_'):
  206. category = pd_name[9:]
  207. if not category in categories:
  208. continue
  209. selected_categories.append(category)
  210. elif pd_name == 'locale' and pd in settings['locales']:
  211. locale = pd
  212. elif pd_name == 'language' and (pd == 'all' or
  213. pd in (x[0] for
  214. x in language_codes)):
  215. lang = pd
  216. resp = make_response(redirect('/'))
  217. if locale:
  218. # cookie max age: 4 weeks
  219. resp.set_cookie(
  220. 'locale', locale,
  221. max_age=60 * 60 * 24 * 7 * 4
  222. )
  223. if lang:
  224. # cookie max age: 4 weeks
  225. resp.set_cookie(
  226. 'language', lang,
  227. max_age=60 * 60 * 24 * 7 * 4
  228. )
  229. if selected_categories:
  230. # cookie max age: 4 weeks
  231. resp.set_cookie(
  232. 'categories', ','.join(selected_categories),
  233. max_age=60 * 60 * 24 * 7 * 4
  234. )
  235. return resp
  236. return render('preferences.html',
  237. locales=settings['locales'],
  238. current_locale=get_locale(),
  239. current_language=lang or 'all',
  240. language_codes=language_codes)
  241. @app.route('/stats', methods=['GET'])
  242. def stats():
  243. global categories
  244. stats = get_engines_stats()
  245. return render('stats.html', stats=stats)
  246. @app.route('/robots.txt', methods=['GET'])
  247. def robots():
  248. return Response("""User-agent: *
  249. Allow: /
  250. Allow: /about
  251. Disallow: /stats
  252. Disallow: /engines
  253. """, mimetype='text/plain')
  254. @app.route('/opensearch.xml', methods=['GET'])
  255. def opensearch():
  256. method = 'post'
  257. # chrome/chromium only supports HTTP GET....
  258. if request.headers.get('User-Agent', '').lower().find('webkit') >= 0:
  259. method = 'get'
  260. ret = render('opensearch.xml', method=method, host=get_base_url())
  261. resp = Response(response=ret,
  262. status=200,
  263. mimetype="application/xml")
  264. return resp
  265. @app.route('/favicon.ico')
  266. def favicon():
  267. return send_from_directory(os.path.join(app.root_path, 'static/img'),
  268. 'favicon.png',
  269. mimetype='image/vnd.microsoft.icon')
  270. def run():
  271. from gevent import monkey
  272. monkey.patch_all()
  273. app.run(
  274. debug=settings['server']['debug'],
  275. use_debugger=settings['server']['debug'],
  276. port=settings['server']['port']
  277. )
  278. if __name__ == "__main__":
  279. run()